Modern marketing teams are swimming in pools of untapped voice data, yet businesses are only scratching the surface of what the technology has to offer in terms of insights and customer knowledge, according to Mike De Cambra.

According to recent estimations, 2.5 quintillion bytes of information are generated every single day, with 90% of the world’s data created in the last two years alone, according to Forbes. But while the statistics are mind-boggling, they’re just the tip of the big data iceberg, with areas like voice data remaining relatively unexplored.

Big data is the term used to describe data sets that are too large or complex for traditional software to process, but also refers to the analysis of such data sets and the extraction of insight and predictions that follow. Advances in technology are allowing more people access to its power and new types of data – such as big voice – are also becoming available.

How many important business phone calls does your company make each year? And how much could you learn about your customer service capabilities from each call recording? With readily available technology now able to interpret speech at a granular level, companies are increasingly able to mine their calls for valuable information – especially in customer services industries.

Why is voice data important?

Studies in the US found that 90% of customers prefer to resolve customer service issues by phone. Each conversation is an untapped opportunity to discover what worked and what didn’t in order to improve the customer experience. By employing voice data analysis, speech intelligence tools can help contact centres map customer sentiment across their call databases, learning how to reduce unnecessary call time, identify dissatisfied customers and analyse employee performance.

Take a telecommunications company, for example. Australia’s five biggest telcos received more than 88 million enquiries from customers dealing with connection or activation issues, billing and plan enquiries, new business, upgrades and plenty more in a one-year period alone between 2017 and 2018. Voice data capture through call recording is most likely already being used to store these calls for legal purposes in case issues arise further down the line, however they are missing out on a honey pot of valuable insights.

In the past and even today, marketing teams have to rely on anecdotal feedback, double jacking on calls or NPS surveys to understand the ‘voice of the customer’. No disrespect to contact centre teams, because I’ve been both in them and managed them (many years ago), but unless they recite the conversation word-for-word, (impossible) they are unable to truly capture a 100% account of actual events rendering the ‘data’ unreliable.

Today, things are changing, and this is where it gets exciting. Aside from basic contact centre analytics providing information like average wait time, the number of transfers, which items were said or selected within the IVR, average handling time and frequency of particular call drivers, big voice analysis allows the telco to unpack its calls to reveal insights on a more sophisticated level.

Insights that big voice technology can uncover

With the evolution of speech intelligence, everyone from enterprises to small businesses and even singular users can access technology to evaluate recorded conversations. This functionality can be used to identify recurring issues and opportunities for business development and can track trends in customer behaviour.

Users can search their recorded calls using a variety of metrics from date and location to specific words, phrases or sentiment. Why is sentiment important? Let’s think about a customer complaint. While the complaint may be ‘resolved’, if the customer sentiment is negative, it doesn’t necessarily mean they are now less of a churn risk than before, it may even be higher.

Sentiment tools can identify the emotions of speakers through a recorded conversation by measuring speech characteristics such as tone of voice, speed of speech, pauses, interruptions and certain phrases that express emotion. These indicators create a detailed understanding of a caller’s emotions throughout the conversation and provide users with a report that assigns the call with a positive, neutral or negative score.

Other tools allow users to track specific words and phrases across their captured communications and jump to the point in a call when the word or phrase was spoken. These words can be tagged to be automatically highlighted in future recordings and automate actions in other applications – opening limitless potential use cases.

If we jump back to the telco example, how fantastic would it be to be alerted when an existing customer mentioned a competitor’s name? Even more exciting is the possibility of triggering an event such as a follow-up call or direct comms to the customer to attempt to truly understand churn risk before it happens.

With such advances already on the market, where can voice data go from here?

The future of big voice

The global speech analytics market is expected to reach $1.6 billion by 2020. While adoption increases for existing tools, the depth of technical capabilities will also improve as the AI and machine learning solutions evolve with every new dataset.

Soon, contact centre agents will be able to measure customer sentiment on the phone in real-time, flagging any severely negative conversations as they happen, so that other staff members can step in to mediate and help the call progress. Callers will also be able to be identified automatically by their voice, saving hours of authentication across the contact centre – an improvement on the innovative but often clunky existing tools that some financial services like the ATO already has in place.

Contact centre agents will also be able to access predictions and trends in their customers’ behaviour, intentions and requirements across the call database. Detailed customer profiles will hold information from all communication channels to track all customer interactions. And insurance and financial fraud detection software may even be able to analyse voice patterns for risk indicators.

All this potential, yet voice data is often excluded from the big data conversation. As we approach the final frontier of digital transformation, and Industry 4.0 extends the layer of smart technology across all industries, marketers should be looking to consolidate voice data in the cloud and capture the valuable insights of important business conversations – straight from the horse’s mouth.

Mike De Cambra is global marketing director at Dubber

Further Reading:

• Have marketers realised the full potential of voice marketing? »
• Forrester VP on voice tech: “Everyone, including Amazon, fails the trust test” »
• Google Duplex changes the game in voice AI – can you tell the difference? »

Image credit:Quino Al